4.9-stable review patch. If anyone has any objections, please let me know. ------------------ From: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx> commit 15cc78644d0075e76d59476a4467e7143860f660 upstream. There was a case that triggered a double free in event_trigger_callback() due to the called reg() function freeing the trigger_data and then it getting freed again by the error return by the caller. The solution there was to up the trigger_data ref count. Code inspection found that event_enable_trigger_func() has the same issue, but is not as easy to trigger (requires harder to trigger failures). It needs to be solved slightly different as it needs more to clean up when the reg() function fails. Link: http://lkml.kernel.org/r/20180725124008.7008e586@xxxxxxxxxxxxxxxxxx Cc: stable@xxxxxxxxxxxxxxx Fixes: 7862ad1846e99 ("tracing: Add 'enable_event' and 'disable_event' event trigger commands") Reivewed-by: Masami Hiramatsu <mhiramat@xxxxxxxxxx> Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- kernel/trace/trace_events_trigger.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) --- a/kernel/trace/trace_events_trigger.c +++ b/kernel/trace/trace_events_trigger.c @@ -1389,6 +1389,9 @@ int event_enable_trigger_func(struct eve goto out; } + /* Up the trigger_data count to make sure nothing frees it on failure */ + event_trigger_init(trigger_ops, trigger_data); + if (trigger) { number = strsep(&trigger, ":"); @@ -1439,6 +1442,7 @@ int event_enable_trigger_func(struct eve goto out_disable; /* Just return zero, not the number of enabled functions */ ret = 0; + event_trigger_free(trigger_ops, trigger_data); out: return ret; @@ -1449,7 +1453,7 @@ int event_enable_trigger_func(struct eve out_free: if (cmd_ops->set_filter) cmd_ops->set_filter(NULL, trigger_data, NULL); - kfree(trigger_data); + event_trigger_free(trigger_ops, trigger_data); kfree(enable_data); goto out; }
